frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 12:54:11 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix c10d -> dist in test_ddp_hooks.py (#61864)

Browse Source 
Summary:
**Overview:**
The existing `test_ddp_hooks.py` test file uses a prefix `c10d`, which is not defined in the file, meaning the test errors if left as is. This renames each `c10d` prefix to `dist`.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/61864

Test Plan:
All four tests pass when run:
```
gpurun python test/distributed/algorithms/ddp_comm_hooks/test_ddp_hooks.py
```

Reviewed By: ejguan

Differential Revision: D29783860

Pulled By: andwgu

fbshipit-source-id: 16bdd2dfcb76192964246148f14851a74f8907c8
This commit is contained in:
Andrew Gu
2021-07-22 07:18:50 -07:00
committed by Facebook GitHub Bot
parent 109bd5e78a
commit 5186fa2831
1 changed files with 8 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
test/distributed/algorithms/ddp_comm_hooks/test_ddp_hooks.py
View File

@ -108,8 +108,8 @@ class DistributedDataParallelCommHookTest(MultiProcessTestCase):
This unit test verifies the ``allreduce`` hook registered case gives same result
with no hook registered case.
"""
store = c10d.FileStore(self.file_name, self.world_size)
process_group = c10d.ProcessGroupNCCL(store, self.rank, self.world_size)
store = dist.FileStore(self.file_name, self.world_size)
process_group = dist.ProcessGroupNCCL(store, self.rank, self.world_size)
# No hook registered case, get the reference grads.
reference_grads = self._get_grads(process_group, None)
@ -125,8 +125,8 @@ class DistributedDataParallelCommHookTest(MultiProcessTestCase):
This unit test verifies the ``fp16 compress`` hook registered case
gives close result with no hook registered case.
"""
store = c10d.FileStore(self.file_name, self.world_size)
process_group = c10d.ProcessGroupNCCL(store, self.rank, self.world_size)
store = dist.FileStore(self.file_name, self.world_size)
process_group = dist.ProcessGroupNCCL(store, self.rank, self.world_size)
# No hook registered case, get the reference grads.
reference_grads = self._get_grads(process_group, None)
@ -142,8 +142,8 @@ class DistributedDataParallelCommHookTest(MultiProcessTestCase):
This unit test verifies the ``quantize per tensor`` hook registered case
gives close result with no hook registered case.
"""
store = c10d.FileStore(self.file_name, self.world_size)
process_group = c10d.ProcessGroupNCCL(store, self.rank, self.world_size)
store = dist.FileStore(self.file_name, self.world_size)
process_group = dist.ProcessGroupNCCL(store, self.rank, self.world_size)
# No hook registered case, get the reference grads.
reference_grads = self._get_grads(process_group, None)
@ -159,8 +159,8 @@ class DistributedDataParallelCommHookTest(MultiProcessTestCase):
This unit test verifies the ``quantize per channel`` hook registered case
gives close result with no hook registered case.
"""
store = c10d.FileStore(self.file_name, self.world_size)
process_group = c10d.ProcessGroupNCCL(store, self.rank, self.world_size)
store = dist.FileStore(self.file_name, self.world_size)
process_group = dist.ProcessGroupNCCL(store, self.rank, self.world_size)
# No hook registered case, get the reference grads.
reference_grads = self._get_grads(process_group, None)